Do Chinese- and English-speaking preschoolers think differently about language?

Abstract

Metalinguistic awareness is the ability to identify, reflect upon, and manipulate linguistic units. It plays a critical role in reading development. The present study investigated Chinese- and English-speaking preschoolers’ metalinguistic awareness development and the role of cognitive and linguistic abilities in its development. Forty-two Chinese-speaking and 36 English-speaking monolingual children completed a series of metalinguistic awareness, false belief, inhibitory control, and receptive vocabulary tasks. The results revealed distinct pathways for the two language groups. English speakers had a more advanced level of rhyme awareness. Chinese speakers developed homonym understanding faster during the preschool years. Inhibitory control was more important for Chinese speakers to develop synonym and homonym understanding, whereas receptive vocabulary was crucial for English speakers to develop rhyme awareness. These differences may be attributable to the characteristics of the Chinese and English languages, as well as the patterns of cognitive development in the two populations.

Keywords

Cross-linguistic comparison false belief inhibitory control metalinguistic awareness preschoolers receptive vocabulary

Language impacts cognition and its development in multiple ways. Some studies have found evidence that child and adult speakers of different languages conceptualize the world differently (Choi & Hattrup, 2012). For example, Mandarin speakers and English speakers have different concepts of time because of differences in productivity and frequency of spatial terms used to talk about time in the two languages (Boroditsky, Fuhrman, & McCormick, 2011). Specifically, vertical spatial terms, /shàng/ (up) and /xià/ (down), are used much more frequently to describe time in Mandarin than in English, leading to Mandarin speakers thinking about time more vertically than English speakers. The current study examines the impact of languages (Chinese vs. English) on metacognitive development, specifically metalinguistic awareness. Metalinguistic awareness is the ability to identify, reflect on, and manipulate linguistic units and is an important predictor of later reading ability (Apel, Wilson-Fowler, Brimo, & Perrin, 2012). During the preschool years, children begin to reflect on the formal properties of language, i.e., thinking about their language (Carroll, Snowling, Stevenson, & Hulme, 2003). This emergence of metalinguistic awareness has been linked to both linguistic (e.g., Silvén, Niemi, & Voeten, 2002) and cognitive (e.g., Doherty & Perner, 1998; Garnham & Garnham, 2002) achievements.

While most of the evidence in this line of research comes from studies of alphabetic languages such as English, less is known about non-alphabetic languages such as Chinese. The present study compared metalinguistic awareness (i.e., phonological and semantic) in English-speaking preschoolers to Mandarin Chinese-speaking preschoolers. We were also interested in whether cognition (i.e., false belief, inhibitory control) and language (i.e., receptive vocabulary) played different roles in metalinguistic development depending on the language spoken.

Cross-linguistic comparisons in metalinguistic awareness

Chinese, as a non-alphabetic language, differs from English in several ways. Such differences in the two languages can lead to distinct pathways for metalinguistic development. For example, in spoken English, syllable boundaries vary depending on levels of stress. Syllables that are stressed do not directly mark stem morphemes even though there is some regularity. Grammatical morphemes, which are not stressed, are shorter because of this pattern. Mandarin, however, does not utilize stress to mark syllables. Instead, tones mark each syllable reflecting a single word (Yeong & Liow, 2012, p. 113). Tone, a suprasegmental phoneme attached to rhyme in each Chinese word, is important in Chinese speech perception, because together with vowels and consonants, tone determines the meaning of a word (Fu, Zeng, Shannon, & Soli, 1998). For example, the same syllable /ma/ with different tones can have different meanings –/mā/ means mom, /má/ means numb, /mǎ/ means horse, and /mà/ means scold. On the other hand, in non-tonal languages such as English, there are no tones attached to rhyme in each word and tones do not change the meaning of the word (Grieser & Kuhl, 1988).

These contrasts in phonology can result in different phonological processing (e.g., rhyme detection) for Chinese and English speech sounds. Studies on alphabetic languages have shown that phonological awareness develops across levels of the syllable during the preschool years (Farrar & Ashwell, 2012; Ho & Bryant, 1997; McBride-Chang, Bialystok, Chong, & Li, 2004). At about 3 or 4 years of age, children are able to perceive and manipulate syllables of a language. This is followed by onset-rhyme awareness, which is the ability to distinguish the onset of a word from the rhyme. For example, the ability to separate cat into /k/ (onset) and /æt/ (rhyme). Rhyme detection, one of the earliest developing phonological skills, emerges around 3 years of age and plays an important role in the acquisition of alphabetic languages (Lonigan, Burgess, & Anthony, 2000).

Cross-linguistic studies indicate that sensitivity to subsyllabic structure (e.g., rhyme) is not as critical to learning Chinese as it is to learning English (Chen, Lin, & Ferrand, 2003). This suggests that in Chinese, phonological awareness may be processed at the syllable level, but phonological awareness in English may be processed as subsyllabic chunks at the phonemic level (see Yeong & Liow, 2012), such as rhyme.

Another important linguistic difference between Chinese and English is that ‘the onset (e.g., h, d, z) of a Chinese syllable is always a single consonant and the rhyme segment (e.g., ai, uo, ao) consists of mainly vowels (Wang & Honig, 2010, p. 142). In contrast, the syllable of English can be very complex, including both singleton (e.g., cat) and cluster onsets (e.g., perspective) (Yang, Yang, & Kang, 2014). That is to say, Chinese is monosyllabic (e.g., mù, yāo) in nature while the English language is multisyllabic (e.g., perspective, university) (Ho & Bryant, 1997). As a result, ‘Mandarin Chinese … has a much smaller number of possible syllables than does spoken English. There are about 400 different syllables in Mandarin Chinese’ (Wang & Honig, 2010, p. 142).

Therefore, because of fewer syllables, homonyms occur more frequently in Chinese than in English (Hannas, 1997), which may lead to distinct semantic processing of the two languages. Semantic awareness includes a variety of abilities, such as understanding homonyms and synonyms. Studies of the development of semantic awareness have focused primarily on alphabetic languages such as English, with few studies on non-alphabetic languages such as Chinese. The understanding of synonyms and homonyms in English emerges around age 4 (Backscheider & Gelman, 1995; Doherty, 2000), although difficulty with both persists until at least age 10 (Beveridge & Marsh, 1991; Corthals, 2010; Mazzocco, 1997).

Linguistic and cognitive Correlates of metalinguistic awareness

In addition to performance differences on metalinguistic tasks, we were interested in the relationship of metalinguistic awareness to preschool children’s other linguistic and cognitive skills (Kamawar & Olson, 2009). There are two major theoretical perspectives regarding these relationships. First, Bialystok and colleagues (Astheimer, Janus, Moreno, & Bialystok, 2014; Bialystok, 1999) have proposed that metalinguistic development reflects linguistic development and is ‘continuous with language learning and use’ (Homer, 2009, p. 489). Empirical studies of alphabetic languages have provided mixed results to this perspective. For example, Dale and colleagues (Crain-Thoreson & Dale, 1992; Dale, Crain-Thoreson, & Robinson, 1995) found no relation between early language development and later phonological awareness, particularly among verbally precocious infants. In contrast, vocabulary development at age 2 predicted performance on rhyming comprehension tasks at age 4 in a two-year longitudinal study (Silvén et al., 2002). No studies have explored these relationships in the Chinese language.

Second, (Hakes, 1980; Sinclair, 1978) have suggested that metalinguistic development is ‘continuous with other, more general cognitive developments’ (Homer, 2009, p. 489). Recent arguments have been more explicit and have suggested that false belief understanding and inhibitory control contribute to phonological and semantic awareness (Tsuji & Doherty, 2014; Yang et al., 2014). Relatedly, studies of other forms of metacognition, such as metamemory, have identified longitudinal relationships with both language and theory of mind in preschoolers (Ebert, 2015).

Metalinguistic awareness and false belief

False belief is a component of theory of mind in which children understand that two individuals have conflicting beliefs regarding the same situation, such as the location of an object. False belief understanding reflects children’s representational flexibility – the ability to handle different representations of the same situation. Similarly, metalinguistic awareness requires children to represent different aspects (e.g., both form and meaning) of language (e.g., Doherty, 2000; Farrar & Ashwell, 2008).

Preschoolers’ understanding of both homonyms and synonyms has been linked to false belief reasoning even after controlling for age and verbal intelligence (Doherty, 2004; Doherty & Perner, 1998). For instance, in a synonym task, children had to understand that the same object could have two words that refer to it, such as sofa–couch; in a homonym task, children had to understand that the same sounding word can have two different meanings, such as flower–flour; and in a false belief task, children had to understand that there can be different beliefs about the same situation (Doherty, 2004).

Similarly, 4-year-olds’ false belief understanding has been linked to rhyme awareness (Farrar & Ashwell, 2008, 2012). As in homonyms and synonyms understanding, both rhyme awareness and false belief require children to consider different representations of the same situation. In a rhyme detection task, children had to select two out of three pictures whose referents – a dog, log, or cat – rhymed. The word dog rhymes with log but has a semantic association with cat. In order to point to the correct answer, dog and log, children had to attend to the sounds of the words while ignoring the meanings of the words. This task required the coordination of two different perspectives: similar sounds and similar meanings.

Metalinguistic awareness and inhibitory control

Inhibitory control also develops during the preschool years and is one component of executive function, which refers to higher-order, self-regulatory cognitive processes used in the monitoring and control of thoughts and actions (Zelazo, Carlson, & Kesek, 2008). Cross-cultural studies have demonstrated that Chinese preschoolers have more advanced levels of executive functioning than their English-speaking counterparts (Sabbagh, Xu, Carlson, Moses, & Lee, 2006). On the other hand, the two groups do not demonstrate such a difference in false belief performance (Sabbagh et al., 2006).

Inhibitory control has also been proposed to relate to metalinguistic development because both require children to inhibit their primary responses (e.g., Farrar & Ashwell, 2012). For instance, in the rhyming task, children had to inhibit the tendency to select the semantic associates and instead choose the phonological associates. However, preschool children’s rhyming ability was significantly related to false belief (Farrar & Ashwell, 2012) but not inhibitory control. (Farrar & Ashwell, 2012). In addition, a Chinese–English bilingual study (Yang et al., 2014) found that Chinese–English bilingual children’s phonological skills were bi-directionally related to overall executive-attentional functioning. Furthermore, the relationship varied by different levels of phonological skill (onset vs. rhyme) and language (English vs. Chinese). Specifically, orienting abilities, which are responsible for selectively allocating attention to stimuli for further processing, were strongly related to onset and rhyme levels in both English and Chinese, whereas executive control abilities, which specialize in control processes, were strongly related to phonological awareness in English only. This suggests that the relation between a bilingual’s phonological awareness skills and executive attention and control might be language-specific. However, inhibitory control was not measured directly in Yang et al.’s (2014) study.

In terms of semantic awareness, Garnham and colleagues (Garnham & Garnham, 2002; Garnham, Brooks, Garnham, & Ostenfeld, 2000) questioned the relationship between false belief and the understanding of synonyms and homonyms found by Doherty (2002). Garnham argued that false belief and homonym tasks did not share a common requirement of representational flexibility. Instead, she proposed an inhibition account of the correlation between false belief and homonym tasks, i.e., the ability to inhibit the most prototypical response was the mechanism underlying the correlation. For example, in a homonym task, when showed a picture labeled flower, children need to inhibit their dominant response to point to a flower in order to point to the correct picture of flour. Similarly, in a unexpected location false belief task, the Sally–Anne test, when asked where would Sally look for her marble after she comes back, children needed to inhibit their tendency to point to Anne’s box in order to correctly point to the less salient location, Sally’s basket. Doherty (2002) rejected this alternative explanation by arguing that Garnham’s homonym task did not specifically require the child to consider simultaneous conflicting representations. However, no studies have directly tested whether inhibitory control is associated with the understanding of homonyms (or synonyms). A direct examination of the relative roles of false belief and inhibitory control in metalinguistic awareness would clarify the roles of cognitive flexibility (Doherty, 2002) and inhibitory control (Garnham & Garnham, 2002). Furthermore, none of these relationships have been explored in Chinese-speaking preschoolers. However, while not assessing metalinguistic awareness, studies of different languages find that inhibitory and language skills are related to false belief reasoning (Cheung et al., 2004; Moriguchi, Okanda, & Itakura, 2008).

Present study

The goals of the present study were to explore metalinguistic development in Chinese- and English-speaking preschoolers and to investigate its cognitive and linguistic correlates. We chose to study preschoolers because it is during the preschool years that children make rapid progress in metalinguistic awareness, executive functioning, and theory of mind. To do this, rhyme, homonym, synonym, false belief, inhibitory control, and receptive vocabulary were examined in the two language populations. It was hypothesized that English speakers would perform better in rhyme detection due to the emphasis on phonemes in English vocabulary, while Chinese speakers would perform better in homonym understanding because of their more advanced inhibitory control ability, as well as the higher frequency of homonyms in Chinese.

Method

Participants

Forty-two Chinese speakers (M = 51.79 months, SD = 9.01 months, range 37–66 months; 21 girls) and 36 English speakers (M = 52.14 months, SD = 6.63 months, range 38–66 months; 20 girls) were recruited from local preschools in China and the United States. Most Chinese children came from middle-class families with at least one parent having a college degree. Most US children came from middle-class families with at least one parent having a college degree or higher. All children were monolingual. Children were divided into two age groups based on a median split; those younger than 54 months were classified into the younger group, and the rest were classified into the older group. Age group ‘37–53 months’ had 40 children (M = 45.35, SD = 4.44); age group ‘54–66 months’ had 38 children (M = 58.89, SD = 3.77).

Procedure

Children were tested individually in their preschools by two experimenters. There were four types of measures: metalinguistic awareness (rhyme, synonym, and homonym), false belief, inhibitory control, and receptive vocabulary. The different tasks for each measure were intermingled across the testing session(s) rather than being presented in a block format. The order of the tasks was the same for all children, with unexpected location as the first task, followed by rhyme detection, day/night stroop, unexpected content, synonym judgment, Peabody Picture Vocabulary Task, dimensional change card sort, appearance–reality, homonym judgment, and head–toes. Children were tested in one or two sessions.

Metalinguistic awareness

Each child completed a rhyme detection task, a synonym judgment task, and a homonym judgment task. Chinese and English stimuli were selected from age-appropriate common words in daily life by the first author and were confirmed by the preschool teachers in order to make sure that the participants have knowledge of the stimuli. The list of experimental sets is given in Appendices A –C.

Rhyme detection

This task was adopted from two previous studies (Farrar, Ashwell, & Maag, 2005; Shu, Peng, & McBride-Chang, 2008). It consisted of two practice trials and 10 experimental trials. Each child was shown 12 sets of three pictures with each picture labeled with a word and then was asked which of the two ‘sound the same’ as the target word. Of the two choices, one rhymed with the target word but had no semantic association and the other did not rhyme but had a semantic association. For example, in English, the target word cat was presented aloud with its picture to children and then followed by hat and chick. In Chinese, the target word /bīng/ (ice) was presented and followed by /dōng/ (winter) and /xīng/ (star). In these cases, the rhyme of hat matched that of cat and the rhyme of /xīng/ matched that of /bīng/. The scores for the rhyme task in both language groups ranged from 0 to 10.

Synonym judgment

The synonym judgment task was adopted from Doherty and Perner (1998). It consisted of three phases: vocabulary check, three practice trials, and four experimental trials. In the synonym judgment task, each child was asked to name a target picture with one of its two names after a vocabulary check of the two names of the target. Then a puppet voiced by the experimenter was asked to say the other name (i.e., the synonym) for the same target with one of the three options: the other name, the same name, and a totally different name. The child’s task was to judge if the puppet correctly said the other name. For example, in English, sofa and couch refer to the same thing. If the child named it sofa and the puppet named it couch, the correct answer in this case was ‘yes’, because the puppet correctly answered the question by using the synonym couch. In Chinese, /gū niáng/ and /nǚ hái/ both mean girl. If the child named it /nǚ hái/ and the puppet also named it /nǚ hái/, the correct answer was ‘no’. This was because the puppet answered the question incorrectly by naming /nǚ hái/ with itself rather than its synonym, /gū niáng/. This rule was made clear to children in practice trials. The scores for this task in both groups ranged from 0 to 4.

Homonym judgment

The homonym judgment task was adopted from Doherty and Perner (1998). It consisted of three phases: vocabulary check, two practice trials, and four experimental trials. In the homonym judgment task, each child was presented with a target picture and four other pictures (target, homonym, two distractors), and then asked to find the picture that has the same name as the target picture. This occurred after the vocabulary check of all the pictures used in the task. For example, in English, the target flower was shown with its picture, and the child’s task was to point out the homonym flour in the four pictures correctly. A Chinese homonym example was /fēng/ (bee) and /fēng/ (wind). The scores for this task in both groups ranged from 0 to 4. A composite metalinguistic awareness score was computed by adding the scores of the three tasks.

False belief

All children completed three false belief tasks, including unexpected location, unexpected content, and appearance–reality. The scores for all three false belief tasks ranged from 0 to 5.

Unexpected location

This task was modeled after the traditional ‘Sally–Anne’ task (Baron-Cohen, Leslie, & Frith, 1985). The scenario was as follows: Tom put the cake into the cupboard in the kitchen and went out. In his absence, his mom took the cake out of the cupboard and put it into the refrigerator. The false belief question was asked following the story: ‘When Tom went back to the kitchen, where would he look for the cake?’ Then children were asked to justify their answers to the false belief question and were asked two memory questions: ‘Where did Tom put the cake?’ and ‘Where is the cake now?’ Children received 1 point if they answered the false belief question and the two memory questions correctly. They received 0 points if they answered any of the three questions incorrectly.

Unexpected content

This task was adopted from Perner, Leekam, and Wimmer (1987). Children were shown a closed crayon box with ribbons inside and were asked what they thought was inside the box before and after (the first false belief question) opening the box. Then a memory question was asked: ‘Can you remember what was inside the box?’ After the memory question, a puppet that had never seen what was inside the box was introduced. A second false belief question was then asked to children: ‘What does the puppet believe is inside the box?’ Children received 2 points if they answered the two false belief questions and the memory question correctly, and received 0 points if they answered the memory question incorrectly even with correct answers for the false belief questions.

Appearance–reality

Each Child was presented with a sponge that looked like a rock and was told: ‘Look what I have! What is it?’ Then the child touched the sponge and was asked: ‘What is it really?’ The first false belief question then followed: ‘When you first saw it before you squeezed it, what did you think it was?’ Then a memory question was asked: ‘Can you remember what it is?’ After the memory question, a puppet that had never touched the sponge was introduced. A second false belief question was then asked to children: ‘What does the puppet think this is, a sponge or a rock?’ Children received 2 points if they answered the two false belief questions and the memory question correctly, and received 0 if they answered the memory question incorrectly even with correct answers in the false belief questions (Gopnik & Astington, 1988).

Inhibitory control

All children completed three inhibitory control tasks including day/night stroop, head–toes, and dimensional change card sort. The scores for all three inhibitory control tasks ranged from 0 to 42.

Day/night stroop

For the day/night stroop task (Gerstadt, Hong, & Diamond, 1994), children were instructed to say ‘day’ when they saw a picture of a moon and ‘night’ when they saw a picture of a sun. The task consisted of four practice trials and 16 experimental trials for both groups. The score ranged from 0 to 16. Reaction time was not recorded.

Head–toes

In the head–toes task (Cameron Ponitz et al., 2008), children were asked to touch their heads when the experimenter said, ‘touch your toes’ and were asked to touch their toes when the experimenter said, ‘touch your head’. The task consisted of four practice trials and 10 experimental trials for both groups. The scores ranged from 0 to 20. Children received 0 points for an incorrect response, 2 for a correct response, and 1 for an initially incorrect and then immediately self-corrected response.

Dimensional change card sort (DCCS)

A shortened version of the DCCS was presented (Frye, Zelazo, & Palfai, 1995). Children were asked to sort four kinds of picture cards: blue/red rabbits and blue/red boats. Children began initially with the color game, sorting cards solely by color and disregarding the pictured shape. Following this, children were told that they were going to play a new game, with the experimenter saying: ‘Now we are going to change to the shape game.’ In the shape game, the rabbits went with the rabbits and the boats went with the boats, regardless of color. Only the score from the shape game was counted. A response was coded as correct if the child sorted each card by the new rule. Scores ranged from 0 to 6.

Receptive vocabulary

The Peabody Picture Vocabulary Task, Fourth Edition (PPVT-IV; Dunn & Dunn, 2007) and PPVT-R (Sang & Miao, 1990) were used to measure the English and Chinese speakers’ receptive vocabulary, respectively. In this task, the child was asked to point out the correct picture among four possibilities. The stop rules for the two groups were the same. The PPVT-IV had a total of 228 items; the PPVT-R for Chinese-speaking children had a total of 175 items and did not have standard scores.

Results

Cross-linguistic comparisons in performance

The first research question focused on cross-linguistic comparisons between Chinese- and English-speaking children regarding each group’s metalinguistic development. Figure 1 illustrate descriptive analyses of children’s performances on all the measures. Table 1 shows raw scores for each task.

Figure 1.

Percent correct of all tasks in the two language groups.

Table 1.

Means and SDs for all measures.

Task	Range	Chinese-speaking children				English-speaking children
		37–53 months (N = 24)		54–66 months (N = 18)		37–53 months (N = 16)		54–66 months (N = 20)
		Mean	SD	Mean	SD	Mean	SD	Mean	SD
Metalinguistic	0–18	9.94	2.91	12.96	1.66	14.25	3.30	15.95	2.21
Rhyme	0–10	4.46	1.82	5.17	1.72	7.94	2.38	8.60	2.06
Synonym	0–4	3.04	1.04	3.89	.32	3.06	.85	3.95	.22
Homonym	0–4	2.44	1.70	3.90	.26	3.25	.93	3.40	1.14
False belief	0–5	.92	1.02	3.00	1.75	1.81	1.28	3.05	1.50
Inhib. control	0–42	32.58	6.90	38.28	3.94	23.00	7.91	32.15	6.56
Day/night	0–16	13.61	3.07	14.47	1.84	9.56	3.79	12.55	2.95
Head–toes	0–20	14.04	5.81	17.63	3.40	8.13	6.58	14.00	5.24
DCCS	0–6	5.00	1.04	5.79	0.63	5.31	0.79	5.60	0.75
PPVT	0–175 (Chinese); 0–228 (English)	55.54	15.81	78.33	19.29	72.56	23.61	102.40	17.35

Notes: Inhibitory control comprised of three tasks: day/night, head–toes, and DCCS. For the English group, the three tasks were not correlated with each other, ps > .14; a two-way random intraclass correlation coefficient (ICC) was computed, ICC (2,3) = .325. For the Chinese group, the only significant correlation was between head–toes and DCCS, r = .32, p = .038; ICC (2,3) = .260.

A series of 2 (language: English; Chinese) × 2 (age: 37–53 months; 54–66 months) ANOVAs on metalinguistic awareness, cognitive, and vocabulary measures were conducted to explore cross-linguistic differences. Specifically, for the composite metalinguistic score, there was an effect of language group, F (1, 78) = 46.70, p < .001, η²_p = .387, and an effect of age, F (1, 78) = 19.50, p < .001, η²_p = .209. English speakers performed better than Chinese speakers. The interaction between the two variables was not significant. For rhyme, there was an effect of language group, F (1, 78) = 57.70, p < .001, η²_p = .438. English speakers performed better than Chinese speakers, but there was not an effect of age or interaction. For homonyms, there was an effect of age group, F (1, 78) = 8.66, p = .004, η²_p = .105, a significant interaction, F (1, 78) = 5.73, p = .019, η²_p = .072, but not an effect of language group. The post-hoc simple effect analyses showed that within the younger age group, English speakers had better performance than Chinese speakers with homonyms, p = .039, while the two language groups did not differ within the older age group. Older children within the Chinese group did better than younger children, whereas the two age groups did not differ in the English group. This interaction indicates that for homonym development, English speakers have better performance at age 3 to 4, but Chinese speakers catch up by age 4 to 5. Finally, for synonyms there was an effect of age group, F (1, 78) = 27.49, p < .001, η²_p = .271, but not an effect of language group, or a significant interaction.

Consistent with previous studies, Chinese children scored higher than their English-speaking counterparts on inhibitory control, F (1, 78) = 27.95, p < .001, η²_p = .274. However, the two language groups did not differ in false belief or receptive vocabulary performance. Older children performed better than younger children in the composite inhibitory control, F (1, 78) = 24.95, p < .001, η²_p = .25, false belief, F (1, 78) = 27.17, p < .001, η²_p = .269, and receptive vocabulary, F (1, 78) = 36.17, p < .001, η²_p = .328. Note that percent correct was used for receptive vocabulary since the two language groups had different items in PPVT. Raw scores were used for the other tasks.

Cognitive and linguistic correlates of metalinguistic awareness

The second research question examined the relationship of cognitive and linguistic abilities to metalinguistic development. Of particular interest was whether false belief, inhibitory control, and receptive vocabulary played different roles in metalinguistic development in the two language groups. Partial correlations controlling for age were initially computed among the measures for the English and Chinese samples, respectively (see Table 2).

Table 2.

Partial correlation (controlling for age) results for Chinese- and English-speaking children.

English
Chinese	1	2	3	4	5	6	7
1. Metalinguistic		.85**	.30	.26	.35*	.25	.55**
2. Rhyme	.65**		.27	−.08	.35*	.15	.44*
3. Synonym	.53**	−.04		.17	.22	.53**	.52**
4. Homonym	.61**	−.06	.47**		−.23	.21	.09
5. False belief	.10	.08	.07	.15		.34*	.23
6. Inhib. control	.44**	.20	.33*	.39*	.24		.47**
7. PPVT	.09	−.001	.10	.10	.46**	.18

Notes: * p < .05, ** p < .01. Percent correct was measured for each task, and the composite score for one measure was computed as the average percent correct of each task that measured it.

Vocabulary and cognition played different roles in metalinguistic development for Chinese- and English-speaking preschoolers. For Chinese-speaking children, both synonym and homonym understanding were related to inhibitory control. To explore this further, we examined the metalinguistic subscale correlations of inhibitory control: only the head–toes task was related to synonym judgment (r = .52, p = .001) and homonym judgment (r = .60, p < .001). For English-speaking children, rhyme detection was correlated with both false belief and receptive vocabulary, and synonym judgment was correlated to both inhibitory control and receptive vocabulary. According to the more detailed subscale correlation analyses of inhibitory control, the head–toes task was related to rhyme (r = .40, p = .018) and synonym (r = .50, p = .003) performance.

To explore these relationships further, data from Chinese and English speakers were combined for a series of hierarchical regressions (see Table 3). Age and language group were entered in step 1, followed by PPVT and language group × PPVT in step 2, head–toes and language group × head–toes in step 3, and false belief and language group × false belief in step 4. The dependent variables were rhyme, homonym, and synonym, respectively. Residuals of all three interactions were used to prevent multicollinearity problems.

Table 3.

Hierarchical regressions in predicting rhyme detection, synonym judgment, and homonym judgment.

Predictor	Rhyme detection				Homonym judgment				Synonym judgment
	B	SE_B	β	ΔR²	B	SE_B	β	ΔR²	B	SE_B	β	ΔR²
Step 1				.47**				.24**				.29**
Age	.03	.03	.10		.08	.02	.47**		.07	.01	.66**
LG	−3.19	.40	−.67**		−.24	.26	−.09		−.13	.14	−.08
Step 2				.08**				.02				.05*
Age	−.01	.03	−.04		.07	.02	.41**		.05	.01	.49**
LG	−3.12	.38	−.66**		−.23	.26	−.09		−.10	.14	−.06
PPVT	4.77	2.23	.23*		1.00	1.54	.09		1.83	.82	.25*
LG × PPVT	−8.86	3.37	−.21**		2.51	2.34	.11		−1.79	1.25	−.12
Step 3				.02				.18**				.11**
Age	−.02	.03	−.07		.04	.02	.26*		.04	.01	.38**
LG	−3.44	.43	−.73**		−.60	.27	−.23*		−.34	.14	−.21*
PPVT	3.10	2.46	.15		−.18	1.52	−.02		.92	.82	.13
LG × PPVT	−6.19	4.36	−.15		−.06	2.69	.00		−2.26	1.44	−.15
Head–toes	.07	.04	.18		.08	.03	.39**		.05	.01	.38**
LG × head-toes	−.04	.08	−.05		.15	.05	.33**		.06	.03	.19*
Step 4				.01				.02				.01
Age	−.03	.03	−.09		.04	.02	.27*		.04	.01	.40**
LG	−3.28	.45	−.69**		−.65	.28	−.25*		−.37	.15	−.22*
PPVT	2.14	2.58	.10		.09	1.60	.01		1.07	.86	.15
LG × PPVT	−6.29	4.85	−.15		−1.03	3.00	−.05		−1.57	1.62	−.11
Head–toes	.06	.04	.15		.08	.03	.40**		.05	.02	.39**
LG × head–toes	−.04	.09	−.05		.14	.05	.32**		.06	.03	.21*
False belief	.20	.16	.14		−.06	.10	−.08		−.03	.05	−.06
LG × false belief	−.09	.31	−.03		.19	.19	.11		−.10	.10	−.09

Note: LG represents language group. * p < .05, ** p < .01.

When rhyme performance was examined, language group was significant through all four steps. PPVT and its interaction with language group were significant when entered in step 2 but became non-significant when cognitive predictors were entered in step 3 and 4. This suggests that cognitive predictors influenced the relation between linguistic predictor and rhyme detection. Post-hoc regressions were conducted in the two language groups separately in order to untangle the interaction between language group and PPVT. The results showed that PPVT was a significant predictor (p = .009) for English speakers’ rhyme performance when entered in step 2, although it became non-significant when head–toes and false belief were entered in steps 3 and 4, respectively. However, PPVT was not a significant predictor for Chinese speakers’ rhyme detection performance. Figure 2 depicts the interaction between language group and PPVT.

Figure 2.

Interaction between language group and PPVT on the rhyme detection task.

When homonym performance was predicted, age and language group were significant. Other important predictors were head–toes and its interaction with language group, which were significant in steps 3 and 4. Post-hoc regressions were conducted in the two language groups separately in order to untangle the interaction between language group and head–toes. The results showed that head–toes was a significant predictor (ps = .001) for Chinese-speaking children’s homonym performance but not for English-speaking children. Figure 3 depicts the interaction between language group and head–toes.

Figure 3.

Interaction between language group and head–toes on the homonym judgment task.

Synonym performance was predicted by age and language group. PPVT was significant when entered in step 2 and became non-significant when cognitive predictors were entered in steps 3 and 4, suggesting its relationship to synonym was influenced by the entered cognitive predictors. Head–toes was significant, and its interaction with language group was marginally significant when entered in step 3. In step 4, when false belief and its interaction with language group were entered, head–toes remained significant and its interaction with language group became significant as well. Post-hoc regressions were conducted in the two language groups separately to examine the language group and head–toes interactions. The results showed that head–toes was a significant predictor (ps = .001) for Chinese-speaking children’s synonym performance, but not for English-speaking children. Figure 4 depicts the interaction between language group and head–toes performance.

Figure 4.

Interaction between language group and head–toes on the synonym judgment task.

Taken together, consistent with the partial correlations, the results from the hierarchical regressions showed that Chinese and English speakers differed in the relationship between metalinguistic awareness and its cognitive and linguistic correlates. Specifically, for the Chinese speakers, inhibitory control was critical for both the synonym and homonym tasks, whereas it was unrelated for the English speakers. On the other hand, receptive vocabulary was important for the English speakers’ rhyme awareness, but not for the Chinese speakers.

Discussion

The present study investigated metalinguistic development and its correlates in Chinese- and English-speaking preschoolers. Findings suggest there are distinct pathways of metalinguistic development in the two populations. English speakers had a more advanced level of rhyme awareness compared to their Chinese-speaking counterparts. The reason for such differences might be due to the distinct linguistic features between English and Chinese. Specifically, the phonemic nature of English makes children more sensitive to perception and manipulation of the sound structures of the language. In contrast, Chinese is perceived at the syllable level, not offering much help for sensitization to the sound structure of the language. Moreover, tone as a suprasegmental feature mainly attached to the rhyme may add extra difficulty for Chinese-speaking children to develop rhyme awareness. Such an assumption was supported by the partial correlation and hierarchical regression results: receptive vocabulary was a stronger predictor than false belief and inhibitory control in rhyme detection, which suggests that linguistic ability played a more important role than cognitive abilities for English speakers.

The results from the partial correlations also suggested that false belief understanding, rather than inhibitory control, played a more important role in rhyme awareness for English speakers. Both the false belief tasks and the rhyme detection task required the ability to simultaneously consider conflicting representations and flexibly shift between differing perspectives of the same situation. However, hierarchical regressions only confirmed the role of receptive vocabulary but not false belief on rhyme awareness, which indicates that false belief is not as strong of a predictor of receptive vocabulary for English speakers’ rhyme awareness. Similar results had been previously obtained for Japanese children (Tsuji & Doherty, 2014). Such results also are in line with those of Read, Macauley, and Furay’s (2014), who reported that there is a significant relation between rhyme and vocabulary retention in English-speaking children, and confirm Farrar et al. (2005) and Silvén et al.’s (2002) findings that vocabulary development is linked to subsequent phonological skills. These results support Bialystok’s (Astheimer et al., 2014; Bialystok, 1999) argument that metalinguistic development reflects linguistic development.

Although there was no language group effect in the homonym task, the significant interaction between language group and age group suggests Chinese speakers develop homonym understanding faster than their English-speaking counterparts during the preschool years. This difference may be attributed to the fact that homonyms occur more frequently in Chinese than in English (Hannas, 1997). The higher frequency of homonyms in the language facilitates Chinese preschoolers’ understanding of homonyms as their cognitive ability develops to a more advanced level. Such assumptions were also supported by the partial correlation and hierarchical regression results: inhibitory control was a stronger predictor than receptive vocabulary, which suggests that cognitive ability rather than linguistic ability was more important for homonym understanding. Specifically, the head–toes task was significantly related to homonym understanding. Such results support the argument that metalinguistic development is linked to general cognitive development (Hakes, 1980; Sinclair, 1978; Tsuji & Doherty, 2014; Yang et al., 2014). It is important to note that the other two inhibitory control tasks were not related to homonym performance. This may be attributable to ceiling effects for the DCCS and day/night in which 75% and 56%, respectively, of Chinese children scored at the top two levels compared to 46% for the head–toes task. Thus, the head–toes task may be a more developmentally sensitive measure of inhibitory control.

For English speakers, the partial correlation results indicated that inhibitory control, but not false belief, was highly related to synonym performance. The results partially supported Garnham and Garnham’s (2002) argument that the relation between false belief and synonym tasks is due to inhibition rather than representational flexibility. Our findings only provided partial support since there was no link found between false belief and synonyms. However, the hierarchical regressions did not find any predictor for semantic awareness in the present study including inhibitory control. Together, these results suggest that receptive vocabulary plays the most important role in metalinguistic development for English-speaking children.

In summary, these findings suggest that metalinguistic development and its components showed distinct pathways in the two language groups. English-speaking children’s metalinguistic development was strongly predicted by linguistic ability, supporting the view that metalinguistic awareness develops as an aspect of language acquisition. In contrast, Chinese-speaking children’s metalinguistic development was strongly predicted by inhibitory control, and thus supports the perspective that metalinguistic awareness is more closely tied to general cognitive development than language acquisition. These distinct pathways between the two populations could be attributable both to the differences in the languages, as well as the associated patterns of cognitive development, particularly inhibitory control. More broadly, these differential patterns indicate that linguistic differences produce variation in metacognition, specifically metalinguistic development.

Limitations, future directions, and implications

A primary limitation of the present study was that it was not a longitudinal study, so the direction of the effects was not clear. Another issue is that because there are unequal items in the Chinese PPVT and English PPVT, and there is no standard PPVT scoring system for Chinese-speaking children, percent correct scores of PPVT instead of standard scores were used in data analyses for both Chinese and English speakers. Note that, unlike standard scores, percent correct could not rule out an age effect in PPVT. Third, the current study only included receptive vocabulary, which is a language measure typically used in previous research (see Milligan, Astington, & Dack, 2007). Having a measure of both expressive and receptive vocabulary would provide a better overall measure of verbal ability. Fourth, it was hard to exclude the possible confounding variable of culture in the current study. The effect of cultural differences (as opposed to linguistic differences) needs to be further explored. For example, would children speaking the same language but from distinct cultural backgrounds think about language differently or not?

Future studies could include additional phonological awareness tasks at different levels, such as a phoneme deletion task and other aspects of executive functioning and language, such as working memory, cognitive flexibility, and grammar. These additional tasks would allow us to explore metalinguistic development and its relation to cognitive and linguistic abilities in a more complete framework. A second future direction could focus on the influence of bilingualism on metalinguistic development and how it would differ from that of monolinguals.

The results of the present study have potential implications for literacy practice and preschool instruction in both Chinese- and English-speaking countries. With the knowledge that language itself could have an impact on children’s cognition of language, educational instructors from different language backgrounds may adapt different educational strategies. Importantly, the present findings point to different routes in different language learning contexts.

Footnotes

Appendix A

Table A1.

Stimuli used in rhyme detection.

English				Chinese
Practice 1	cat	hat	chick	rì (sun)	dì (earth as in the ground)	yuè (moon)
Practice 2	frog	turtle	fog	bīng (ice)	dōng (winter)	xīng (star)
1.	cake	knife	lake	huǒ (fire)	guǒ (fruit)	shuǐ (water)
2.	sock	shoe	lock	kǒu (mouth)	shǒu (hand)	ěr (ear)
3.	pear	bear	apple	jú (chrysanthemum)	zhú (bamboo)	méi (plum blossom)
4.	rice	ice	soup	huā (flowers)	chūn (spring)	guā (watermelon)
5.	hand	sand	teeth	fàn (meal)	ròu (steak)	dàn (raw eggs)
6.	beef	egg	leaf	yá (teeth)	wá (baby)	bí (nose)
7.	spoon	moon	fork	xiāng (suitcase)	bāo (handbag)	qiāng (toy gun)
8.	peach	lemon	beach	yǐ (ant)	niǎo (bird)	mǐ (rice)
9.	eye	sky	mouth	kù (pants)	tù (rabbit)	xié (shoes)
10.	shirt	pants	desert	wā (frog)	guī (turtle)	shā (sand)

Appendix B

Table B1.

Stimuli used in synonym judgment.

English		Chinese
Practice 1 ship	boat	xiǎo zhōu (boat)	xiǎo chuán (boat)
Practice 2 cup	mug	xiǎo péng yǒu (child)	xiǎo hái (child)
Practice 3 coat	jacket	gāo lóu (tall building)	dà shà (tall building)
1. sofa	couch	xiǎo niǎo (little bird)	xiǎo yàn zi (little bird)
2. rabbit	bunny	fáng zi (cabin)	mù wū (cabin)
3. present	gift	lán zi (basket)	kuāng (basket)
4. TV	television	nǚ hái (girl)	gū niáng (girl)

Appendix C

Table C1.

Stimuli used in homonym judgment.

English		Chinese
Practice 1 flower	flour	zhū (pig)	zhū (pearl)
Practice 2 letter (envelope)	letter (words)	jī (plane)	jī (chick)
1. nail (finger)	nail (tool)	fēng (wind)	fēng (bee)
2. bat (baseball)	bat (flying)	hóng (red)	hóng (rainbow)
3. son	sun	wǎn (night)	wǎn (bowl)
4. pear	pair	lù (road)	lù (deer)

Acknowledgements

We thank all the children, teachers, and volunteers who facilitated this research and W. Keith Berg, Andreas Keil, and Tian Lin for helpful feedback.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Apel

Wilson-Fowler

E. B.

Brimo

Perrin

N. A.

(2012). Metalinguistic contributions to reading and spelling in second and third grade students. Reading and Writing, 25, 1283–1305. doi:10.1007/s11145-011-9317-8

Astheimer

Janus

Moreno

Bialystok

(2014). Electrophysiological measures of attention during speech perception predict metalinguistic skills in children. Developmental Cognitive Neuroscience, 7, 1–12. doi:10.1016/j.dcn.2013.10.005

Backscheider

A. G.

Gelman

S. A.

(1995). Children’s understanding of homonyms. Journal of Child Language, 22, 107–127. doi:10.1017/S030500090000965X

Baron-Cohen

Leslie

A. M.

Frith

(1985). Does the autistic child have a ‘theory of mind’? Cognition, 21, 37–46. doi:10.1016/0010-0277(85)90022-8

Beveridge

Marsh

(1991). The influence of linguistic context on young children’s understanding of homophonic words. Journal of Child Language, 18, 459–467. doi:10.1017/S0305000900011168

Bialystok

(1999). Cognitive complexity and attentional control in the bilingual mind. Child Development, 70, 636–644. doi:10.1111/1467-8624.00046

Boroditsky

Fuhrman

McCormick

(2011). Do English and Mandarin speakers think about time differently? Cognition, 118, 126–132. doi:10.1016/j.cognition.2010.09.010

Cameron Ponitz

C. E.

McClelland

M. M.

Jewkes

A. M.

Connor

C. M.

Farris

C. L.

Morrison

F. J.

(2008). Touch your toes! Developing a direct measure of behavioral regulation in early childhood. Early Childhood Research Quarterly, 23, 141–158. doi:10.1016/j.ecresq.2007.01.004

Carroll

J. M.

Snowling

M. J.

Stevenson

Hulme

(2003). The development of phonological awareness in preschool children. Developmental Psychology, 39, 913–923.

10.

Chen

Lin

Ferrand

(2003). Masked priming of the syllable in Mandarin Chinese speech production. Chinese Journal of Psychology, 45, 107–120.

11.

Cheung

Hsuan-Chih

Creed

Ping Wang

(2004). Relative roles of general and complementation language in theory-of-mind development: Evidence from Cantonese and English. Child Development, 75, 1155–1170.

12.

Choi

Hattrup

(2012). Relative contribution of perception/cognition and language on spatial categorization. Cognitive Science, 36, 102–129. doi:10.1111/j.1551-6709.2011.01201.x

13.

Corthals

(2010). Short report: Nine- to twelve-year olds’ metalinguistic awareness of homonymy. International Journal of Language & Communication Disorders, 45, 121–128. doi:10.3109/13682820902745446

14.

Crain-Thoreson

Dale

P. S.

(1992). Do early talkers become early readers? Linguistic precocity, preschool language, and emergent literacy. Developmental Psychology, 28, 421–429. doi:10.1037/0012-1649.28.3.421

15.

Dale

P. S.

Crain-Thoreson

Robinson

N. M.

(1995). Linguistic precocity and the development of reading: The role of extralinguistic factors. Applied Psycholinguistics, 16, 173–187. doi:10.1017/S0142716400007074

16.

Doherty

M. J.

(2000). Children’s understanding of homonymy: Metalinguistic awareness and false belief. Journal of Child Language, 27, 367–392. doi:10.1017/S03050009-00004153

17.

Doherty

M. J.

(2002). Synonyms, homonyms and false beliefs: A reply to Garnham, Brooks, Garnham and Ostenfeld. Developmental Science, 5, 84–85. doi:10.1111/1467-7687.00212

18.

Doherty

M. J.

(2004). Children’s difficulty in learning homonyms. Journal of Child Language, 31, 203–214. doi:10.1017/S030500090300583X

19.

Doherty

M. J.

Perner

(1998). Metalinguistic awareness and theory of mind: Just two words for the same thing? Cognitive Development, 13, 279–305. doi:10.1016/S0885-2014(98)90012-0

20.

Dunn

L. M.

Dunn

L. M.

(2007). Peabody Picture Vocabulary Test: Manual. Upper Saddle River, NJ: Pearson Education.

21.

Ebert

(2015). Longitudinal relations between theory of mind and metacognition and the impact of language. Journal of Cognition and Development, 16, 559–586.

22.

Farrar

M. J.

Ashwell

(2008). The role of representational ability in the development of phonological awareness in preschool children. In Farrar

M. J.

Ashwell

(Eds.), Literacy processes: Cognitive flexibility in learning and teaching (pp. 71–89). New York, NY: Guilford Press.

23.

Farrar

M. J.

Ashwell

(2012). Phonological awareness, executive functioning, and theory of mind. Cognitive Development, 27, 77–89. doi:10.1016/j.cogdev.2011.08.002

24.

Farrar

M. J.

Ashwell

Maag

(2005). The emergence of phonological awareness: Connections to language and theory of mind development. First Language, 25, 157–172. doi:10.1177/0142723705051980

25.

Frye

Zelazo

P. D.

Palfai

(1995). Theory of mind and rule-based reasoning. Cognitive Development, 10, 483–527. doi:10.1016/0885-2014(95)90024-1

26.

Zeng

Shannon

R. V.

Soli

S. D.

(1998). Importance of tonal envelope cues in Chinese speech recognition. Journal of the Acoustical Society of America, 104, 505–510. doi:10.1121/1.423251

27.

Garnham

W. A.

(2002). Metarepresentation or inhibition? An open question: A response to Doherty. Developmental Science, 5, 86. doi:10.1111/1467-7687.00213

28.

Garnham

W. A.

Brooks

Garnham

Ostenfeld

(2000). From synonyms to homonyms: Exploring the role of metarepresentation in language understanding. Developmental Science, 3, 428–441. doi:10.1111/1467-7687.00137

29.

Gerstadt

C. L.

Hong

Y. J.

Diamond

(1994). The relationship between cognition and action: Performance of children 3.5–7 years old on a stroop-like day-night test. Cognition, 53, 129–153. doi:10.1016/0010-0277(94)90068-X

30.

Gopnik

Astington

J. W.

(1988). Children’s understanding of representational change and its relation to the understanding of false belief and the appearance-reality distinction. Child Development, 59, 26–37.

31.

Grieser

D. L.

Kuhl

P. K.

(1988). Maternal speech to infants in a tonal language: Support for universal prosodic features in motherese. Developmental Psychology, 24, 14–20. doi:10.1037/0012-1649.24.1.14

32.

Hakes

D. H.

(1980). The development of metalinguistic abilities in children. New York, NY: Springer.

33.

Hannas

W. C.

(1997). Asia’s orthographic dilemma. Honolulu: University of Hawaii Press.

34.

C. S.

Bryant

(1997). Development of phonological awareness of Chinese children in Hong Kong. Journal of Psycholinguistic Research, 26, 109–126. doi:10.1023/A:1025016322316

35.

Homer

B. D.

(2009). Literacy and metalinguistic development. In Olson

D. R.

Torrance

(Eds.), The Cambridge handbook of literacy (pp. 487–500). Cambridge, UK: Cambridge University Press.

36.

Kamawar

Olson

D. R.

(2009). Children’s understanding of referentially opaque contexts: The role of metarepresentational and metalinguistic ability. Journal of Cognition and Development, 10, 285–305. doi:10.1080/15248370903389499

37.

Lonigan

C. J.

Burgess

S. R.

Anthony

J. L.

(2000). Development of emergent literacy and early reading skills in preschool children: Evidence from a latent-variable longitudinal study. Developmental Psychology, 36, 596–613. doi:10.1037/00121649.36.5.596

38.

Mazzocco

M. M.

(1997). Children’s interpretations of homonyms: A developmental study. Journal of Child Language, 24, 441–467. doi:10.1017/S0305000997003103

39.

McBride-Chang

Bialystok

Chong

K. K.

(2004). Levels of phonological awareness in three cultures. Journal of Experimental Child Psychology, 89, 93–111. doi:10.1016/j.jecp.2004.05.001

40.

Milligan

Astington

J. W.

Dack

L. A.

(2007). Language and theory of mind: Meta-analysis of the relation between language ability and false-belief understanding. Child Development, 78, 622–646.

41.

Moriguchi

Okanda

Itakura

(2008). Young children’s yes bias: How does it relate to verbal ability, inhibitory control, and theory of mind? First Language, 28, 431–442.

42.

Perner

Leekam

S. R.

Wimmer

(1987). Three-year-olds’ difficulty with false belief: The case for a conceptual deficit. British Journal of Developmental Psychology, 5, 125–137.

43.

Read

Macauley

Furay

(2014). The Seuss boost: Rhyme helps children retain words from shared storybook reading. First Language, 34, 354–371.

44.

Sabbagh

M. A.

Carlson

S. M.

Moses

L. J.

Lee

(2006). The development of executive functioning and theory of mind. A comparison of Chinese and U.S. preschoolers. Psychological Science, 17, 74–81. doi:10.1111/j.1467-9280.2005.01667.x

45.

Sang

Miao

(1990). The revision of trail norm of Peabody Picture Vocabulary Test Revised (PPVT-R) in Shanghai proper. Psychological Science, 5, 20–25.

46.

Shu

Peng

McBride-Chang

(2008). Phonological awareness in young Chinese children. Developmental Science, 11, 171–181. doi:10.1111/j.1467-7687.2007.00654.x

47.

Silvén

Niemi

Voeten

M. J.

(2002). Do maternal interaction and early language predict phonological awareness in 3-to 4-year-olds? Cognitive Development, 17, 1133–1155. doi:10.1016/S0885-2014(02)00093-X

48.

Sinclair

(1978). Conceptualization and awareness in Piaget’s theory and its relevance to the child’s conception of language. In Sinclair

Jarvella

R. J.

Levelt

W. J. M.

(Eds.), The child’s conception of language (pp. 191–200). New York, NY: Springer.

49.

Tsuji

Doherty

M. J.

(2014). Early development of metalinguistic awareness in Japanese: Evidence from pragmatic and phonological aspects of language. First Language, 34, 273–290. doi:10.1177/0142723714538003

50.

Wang

H. H.

Honig

A. S.

(2010). What difficulties do children experience while learning to read and write Chinese? In Chen

Wang

Cai

(Eds.), Teaching and learning Chinese: Issues and perspectives (pp. 139–154). New York, NY: Information Age.

51.

Yang

Kang

(2014). The relationship between phonological awareness and executive attention in Chinese–English bilingual children. Cognitive Development, 30, 65–80. doi:10.1016/j.cogdev.2013.11.003

52.

Yeong

S. H.

Liow

S. R.

(2012). Development of phonological awareness in English–Mandarin bilinguals: A comparison of English–L1 and Mandarin–L1 kindergarten children. Journal of Experimental Child Psychology, 112, 111–126. doi:10.1016/j.jecp.2011.12.006

53.

Zelazo

P. D.

Carlson

S. M.

Kesek

(2008). The development of executive function in childhood. In Nelson

C. A.

Luciana

(Eds.), Handbook of developmental cognitive neuroscience (2nd ed., pp. 553–574). Cambridge, MA: MIT Press.